The FCC granted an extension today on the report on the risks of LightSquared's proposed broadband data network interfering with GPS receivers. It is clear that the system would cause some interference. We need to quantify how much and what the design tradeoffs are.
I contacted LightSquared and the Coalition to Save Our GPS (CSOG) yesterday morning with technical questions. The PR people at LightSquared seemed to want to help but didn't have immediate access to technical details. CSOG has not responded yet.
In the absence of detailed information, as engineers we can at least go to communications theory to work out the tradeoffs of locating this system so close to the GPS band.
GPS band: 1559-1610MHz
LightSquared satellite uplink band: 1525-1559.
LightSquared satellite downlink band: 1626.5-1660.5 MHz
GPS signals at the earth’s surface have a received power of around -130dBm. GPS receivers need to receive signals as weak as -150dBm and below.
The critical question is how strong will the signals from LightSquared’s network be. Their uplinks would be allowed a maximum of 42dBW (15800W), however, they have said they will limit it to 32 dBW (1580W) EIRP near the GPS band border. The mobile satellite service (MSS) portion of the system would use highly directional “spot beams” covering an area of the earth about 400 to 500km across. They would reuse frequencies in cell fashion so that adjacent "spot beams" could overlap on different frequencies and frequencies would be reused in non-adjacent cells. The ancillary terrestrial component (ATC) of the system would use the same frequencies and in some locations intentionally overpower the satellite signal. (See this FCC document for further reading.)
Important Questions on the Details of LightSquared's System
- We know the EIRP limits on the satellite uplink portion. What is the maximum output power?
- Are there different rules governing the ATC portion? How directional will the typical ATC station’s antenna be?
- Althought it’s likely moot to the question of GPS interference, it would be nice to know the received power of MSS downlink signals. (The orbital spot-beam system sounds cool.)
Calculating Interferer Signal Strength
It is simple matter to work out the strength of a signal transmitted with a given EIRP at given distance. Free Space Path Loss (FSPL) [dB] = 20*log(4*pi*d/wavelength). A more convenient way to express this is FSPL [dB] = 20*log(d[m]) + 20*log(f[MHz) - 27.55.
FSPL(100m,1550MHz) = 76dB.
FSPL(1000m, 1550MHz) = 96dB.
We know the maximum uplink EIRP near the GPS band is 62dBm. So 100m away, the received signal strength would be 62dBm - 76dB = -14dBm. At 1000m distance, it would be -34dBm. The 62dBm EIRP transmitter will most likely be something like a 30dBm [1W] transmitter into a 32dB gain dish. The beam of strong signal coming from such a dish would only be about 5 degrees wide. Outside that narrow cone, the signal will drop around 50dB. The trouble here is the proposed MSS system would involve tens of thousands of such uplinks, each with its own 5-degree-wide cone.
Tradeoffs
- Policy makes must decide how close from a LightSquared transmitter we want GPS receivers to work. No one expects a GPS receiver to work well less than a meter from a transmitter, even if that transmitter is a something like a police radio on a distance frequency band. OTOH, no one wants something that disrupts GPS units hundreds of meters away in all directions.
- GPS manufactures need to work out a curve with real units for tolerance to adjacent band interference (in dBm) versus added cost (in dollars).
Conclusion
The issue of adjacent band interference is nothing new. It’s a constant issue for anyone designing receiving equipment that needs to work near paging towers. Amateur radio operators complained about it when the FCC reduced the size of the 220MHz band, forcing weak-signal work to share a smaller band with more powerful FM repeaters. One thing that is different in the case of GPS is that we are very dependent on a system that relies on signals of -140dBm signal strength. We should consider some sort of supplement (perhaps terrestrially-based) to the GPS system. Many readers will immediately think, “Great idea! We could call it LORAN.” I am not saying whether such a system should be at all similar to LORAN. I’m just saying we could benefit from a system that works on somewhat stronger signals.
Given the GPS system we have today, we need to first answer the questions about acceptable intereference range and acceptable added GPS receiver cost before we can say whether putting LightSquared’s proposed network so close to the GPS band is reasonable. Given the strength of the the proposed network's system, it seems impossible to adopt it without adding some cost to GPS receivers, which could be in some way subsidized by LightSquared subscribers.
